J. Comp. Path. 1997 Vol. 117, 1-16

Experimental Infection of Non-pregnant and Pregnant Sheep with Neospora caninum

D. Buxton, S. W. Maley, K. M. T h o m s o n , A.J. Trees* and E. A. Innes

Moredun Research Institute, 408 Gilmerton Road, Edinburgh EH17 7o7H and *Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, UK

=j ( F

S u m l T l a r ~

In an initial experiment, 21 sheep in groups of five or six were inoculated subcutaneously (sc) with l 08, 106 or 104 Neospora caninum tachyzoites (Liverpool isolate), or with control inoculum, and monitored for clinical signs and for "seroconversion". Animals given the two higher doses showed febrile responses and all three groups inoculated with the parasite showed seroconversion. In a second experiment, 12 pregnant sheep were each inoculated sc at 90 days' gestation with 106 tachyzoites, and at 25, 40 and 53 days post-inoculation (dpi) groups of four were killed for examination of the fetuses and placentas. Appropriate control ewes were included in the study. All fetuses were alive immediately before their dams were killed, except for one, which was found to be mummified at 40 dpi. Histopathological lesions were found consistently in both fetal central nervous system (CNS) and placental tissues. In the latter, focal necrosis, which was mild at 25 dpi, was much more severe at 40 dpi and much less severe at 53 dpi. Lesions in the fetal CNS consisted of focal microgliosis (with or without central necrosis), lymphoid cuffing and non- suppurative meningitis. Lesions were also found in fetal liver, heart and lung. Neospora antigen was demonstrated in fetal brain and placental tissues and, at 25 dpi, in single samples of fetal liver and heart. The prescapular lymph nodes did not differ in size from those of control fetuses but were more mature in that they contained a significantly greater number of secondary follicles. Both IgM and IgG antibodies to N. caninum were detected in the serum of fetuses from infected ewes. Thus, N. caninum readily infected pregnant ewes and caused lesions in fetal tissues and placentas which resembled those of ovine toxoplasmosis. In addition, the changes were similar to those of bovine neosporosis; the infected pregnant ewe therefore offers a good model for the bovine disease. �9 1997 W.B. Saunders Company Limited

Introduction

T h e p ro tozoan parasi te known as Neospora caninum was first r eco rded in 1984, as the cause o f encephali t is ~md myositis in six dogs in N o r w a y (Bjerkas et al., 1984). In 1988 it was isolated and n a m e d by D u b e y et al. (1988) and is now recognized as a cause of abor t ion and neona ta l loss in cattle (Thilsted and D u b e y , 1989; Bar r et al., 1990; T h o r n t o n et al., 1991; W o u d a et al., 1992), horses (Dubey and Porterfield, 1990), goats (Barr et al., 1992; D u b e y et al., 1992, 1996) and sheep (Dubey et al., 1990); it has also been repor ted as the cause o f acute illness and dea th in a black-tai led deer (Odocoileus hemionus)

0021 9975/97/050001 § 16 $12.00/0 �9 1997 W.B. Saunders Company Limited

2 D. Buxton e t al.

(Woods et al., 1994). Neospora caninum is closely related to Toxoplasma gondii (Ellis et al., 1994; H ohndah l et al., 1994) and, in a pre l iminary report, Buxton et al. (1995) noted that experimental infection of pregnant sheep with N. caninum induced disease, with lesions similar to those seen in ovine toxoplasmosis (Buxt0n and Finlayson, 1986). Such lesions were also described by McAllister et al. (1996), who showed that, as in ovine toxoplasmosis, the stage of gestation appeared to influence the outcome of infection.

This paper records an initial experiment in male sheep given different doses of N. caninum tachyzoites to establish an infective dose and clinical and serological parameters . In a second experiment, the pathogenici ty of the parasite for pregnant sheep and their fetuses was examined over a period of time.

Mater ia l s and M e t h o d s

Animals and Treatments

Experiment 1. Twenty-one male, castrated, year-old Scottish Blackface sheep were used. The animals, which were seronegative for T. gondii by an enzyme-linked immunosorbent assay (ELISA) for anti-toxoplasma IgG (Buxton et al., 1988) and seronegative for N. caninum by an indirect fluorescent antibody test (IFAT; see below), were divided into one group of six sheep (group 1.1) and three groups of five (groups 1.2, 1.3 and 1.4). The sheep in groups 1.1, 1.2 and 1.3 received doses of 108, 106and 104N. caninum tachyzoites, respectively, in a 2-ml inoculum injected subcutaneously (sc) over the left prefemoral lymph node. The same volume of control inoculum was administered similarly to sheep in group 1.4.

Experiment 2. Twelve pregnant ewes with synchronized mating dates (Shetland cross Scottish Blackface), aged 4-5 years, were used as group 2.1. They were seronegative for T. gondii by the ELISA and for N. caninum by the IFAT. At 90 days' gestation, each received 106 N. caninum tachyzoites (dose volume 2 ml) sc over the left prefemoral lymph node. Group 2.2 consisted of eight ewes given 2 ml of control inoculum.

Inoculum

N. caninum tachyzoites of the Liverpool isolate (Barber et al., 1995) were cultured in Vero cells, grown in RPMI 1640 (Gibco, Paisley) containing fetal bovine serum (FBS) 2%, penicillin 100U/ml and streptomycin 100gg/ml (Northumbria Biologicals, Cramlington), in plastic culture flasks with a 25 c m 2 surface (Bibby Sterilin Ltd, Stone). Infected cultures were harvested after 7 days by scraping the infected Vero cells from the inner surface of the flasks and centrifuging them at 800 g for 15 min at 4~ The extracellular tachyzoites were then resuspended in the same culture medium and their numbers adjusted to provide the required concentration. In experiment 1, suspensions containing 5 x 10 7, 5 • 10 5 and 5 • 103 tachyzoites/ml were prepared. The suspension containing 5 x 107 tachyzoites/ml was contaminated with Vero cells (1"7 x 107/ml), which were non-viable as indicated by a dye exclusion test. Control inoculum for experiment 1 also contained 1.7 x 10 7 non-viable Vero cells/ml. In experiment 2, the number of tachyzoites was adjusted to 5 x 105/ml and contamination by Vero cells (all non-viable) was estimated to be 2"5 • 105/ml. Control inoculum for experiment 2 also contained 2"5 x 105 non-viable Vero cells/ml.

Clinical Monitoring and Sampling

Rectal temperatures of all sheep, in both experiments, were recorded daily until day 15 after inoculation. In experiment 1, blood samples were collected weekly for

Infection o f Sheep with Neospora eaninum 3

6 weeks from the day of infection and thereafter less frequently for up to 1 year. In experiment 2, blood was collected weekly from the day of infection until necropsy, 25, 40 or 53 days later, when a final sample was obtained. In all cases the serum was separated and stored frozen at --20~ until examined by the IFAT for IgG antibody to N. caninum.

Ser0/0~ In experiments 1 and 2, IgG antibody to N. caninum was sought in all sheep serum samples with an IFAT as outlined below. In experiment 2, fetal serum samples and cerebrospinal fluid (CSF) were also examined for IgG and IgM antibodies to the parasite by the IFAT. For the detection of IgG antibody to N. caninum, tachyzoites of the NC1 isolate (Lindsay and Dubey, 1989) were grown in Vero cells in Iscove's modified Dulbecco's medium (IMDM) (Gibco, Paisley) containing horse serum 2% (free from detectable antibodies to N. caninum), penicillin 100 units/ml and streptomycin 100gg/ml. After three rinses in phosphate-buffered saline (PBS; 0"14MNaC1, 3 mMKC1, 10mMNa2HPO4 and 2mMKH~PO4, pH 7"2) tachyzoites were re- suspended in PBS at a concentration of 107/ml and treated for 24h at 4~ with formaldehyde (40%) at a final concentration of 0-2% v/v. Aliquots were then stored at -- 20~ When required, they were thawed and 5 gl of suspension were placed in each well of a Multitest 15-well slide (ICN Biomedicals Inc., Ohio) and allowed to dry. Doubling dilutions (starting at 1 in 16) of each test serum were added to the wells, which, after incubation at room temperature for 30 min, were thoroughly rinsed with PBS before being flooded with donkey anti-ovine IgG serum labelled with fluorescein isothiocyanate (FITC) (Sigma-Aldrich Company Ltd, Poole). In experiment 2, IgM antibody to N. caninum was detected in fetal serum samples by replacing the donkey anti-IgG conjugate with rabbit antibody to ovine IgM labelled with FITC (Kirkegaard and Perry Laboratories Inc., Gaithersburg). Known standard positive and negative sera were included on all slides to ensure consistency. The slides were viewed under an Olympus BX50 fluorescence microscope, with a U-MNB filter cube, and a x 40 objective.

Necropsy

Only animals in experiment 2 were submitted to a post-mortem examination. Just before necropsy, a sample of blood was collected from each ewe to provide serum. In group 2.1, four ewes were killed with intravenous pentobarbitone sodium (Euthatal; Rhone Merieux Ltd, Harlow) on day 25, 40 or 53 after inoculation. In group 2.2 (controls), two ewes were killed at 25 days, two at 40 days and four at 53 days. On each day of necropsy, the control animals were killed before the infected ewes. Immediately after death, CSF was obtained from the cerebellomedullary cistern and each ewe was hysterectomized through a midline incision. The uterus was opened and, with aseptic precautions, a sample of amniotic fluid was collected. Blood was taken from the umbilical artery and CSF from the cerebellomedullary cistern of each fetus. The uterus and placenta were perfused simultaneously through the middle uterine and umbilical arteries (Barlow, 1972), each with a minimum of 500 ml of 10% formol saline, and subsequently immersed in the same fixative. After perfusion- fixation, fetal and placental tissues were further dissected for histological and immuno- histochemical examination.

Histopathology

Each fetal brain was cut coronally, and slices selected for study included: frontal cerebrum and striatum at the level of the caudate nucleus; parietal cerebrum and thalamus at the level of the optic chiasma; occipital cerebrum and hippocampus; midbrain at the level of the rostral colliculi; medulla at the level of the obex and the

4 D. Buxton e t ai.

cerebellar peduncles; and sagittal cerebellum. Transverse and longitudinal blocks were taken from the cervical and sacral enlargements of the spinal cord, and from the thoracic section. In addition, samples were taken from the placentome (three blocks), the left prescapular lymph node (at its widest part), mesenteric lymph node, spleen, liver, lung, heart, kidney and intestines. All blocks of tissue were fixed in 10% formol saline for 48 h, dehydrated through graded alcohols, treated with toluene, and embedded in paraffin wax. Sections (5 Ixm) were stained with haematoxylin and eosin (HE) and some were stained by the Martius scarlet blue (MSB) method for fibrin, and Gordon and Sweet's method for reticulin. From each ewe, blocks from the brain (similar to those from the fetuses), spleen and left prefemoral lymph node were also collected and processed to paraffin wax; sections were then stained with HE.

Immunohistochemist~y

For the demonstration of N. caninum antigen in experiment 2, selected paraffin-wax sections, 5 gm thick, were cut from fetal brain, heart and liver, and from all blocks of prescapular lymph node and placentome. They were dewaxed with xylene and treated for 30 rain at room temperature with normal goat serum diluted 1 in-4 with 0"5 M sodium chloride in 0-01 M phosphate buffer pH 7"5 (diluent buffer). This was removed by blotting; polyclonal rabbit antiserum to N. caninum (NC 1) was then applied at a dilution of 1 in 1000 in diluent buffer and left at 4~ overnight. The antiserum had been obtained from a rabbit 8 weeks after injecting 7 x 1 0 7 live tachyzoites sc. After incubation, slides were rinsed with 0"5 M sodium chloride in 0"01 M phosphate buffer pH 7-5 (wash buffer), treated with biotinylated goat-anti-rabbit IgG for 30 min at room temperature, and then rinsed in wash buffer. Endogenous peroxidase was blocked with hydrogen peroxide 1% in methanol for 30 min at room temperature; the slides were then rinsed in wash buffer and treated with avidin-biotin complex (ABC) (Vectastain Elite ABC kit; Vector Laboratories, Peterborough). The reaction was developed with diaminobenzidine tetrahydrochloride (DAB) (Sigma Fast, Sigma- Aldrich Company Ltd, Poole, England) for 2 to 5 min. Sections were washed in water, counterstained with haematoxylin and mounted under coverslips. Sections from control animals were prepared for each batch of sections treated in this way. Further controls included in each batch consisted of sections treated as above but omitting (1) antiserum to the parasite and replacing it with normal rabbit serum, or (2) second serum (biotinylated goat anti-rabbit IgG), or (3) both these reagents and the avidin and biotinylated horseradish peroxidase complex (i.e., treating only with DAB). In addition, a positive control section from a case of canine cerebral neosporosis was treated with the full ABC technique as described above.

Quantification of Placental Lesions

Sections (5 gm) were prepared from each of three placentomes from each fetus and stained with HE. The necrotic foci on each section were then counted. The complete area of each section was viewed under an Olympus BX50 microscope fitted with a Sony DXC-107P video camera. The area of the section was obtained by electronic capture of the video image and analysis of a grey scale 512 x 512 pixel image with Semper 6 Plus image analysis software for Windows (Synoptics Ltd, Cambridge). Areas were expressed a s m m 2 by reference to standard areas previously measured at the same magnification. The numbers of follicles in all sections of fetal prescapular lymph nodes stained by Gordon and Sweet's method were also counted.

Statistical Ana~sis

Fetal antibody titres were compared by means of a Mann-Whitney two-sample test and the one-tailed P value was calculated. Lymph node follicles in the prescapular lymph nodes were compared by means of a two-tailed t-test.

I n f e c t i o n o f S h e e p w i t h Neospora caninum 5

41.5

?

41.0

40.5

40.0

39.5

39.0

! k

; ~-- / ',\,

, i ,. / . .

/ ,, . / . , ,

.. " , ) . . . . . . . .

I $ I I I I I I I I I I I I I 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

D a y s a f t e r i nocu l a t i on

Fig. 1. Mean rectal temperatures of the four groups of sheep in experiment 1 given subcutaneous doses of 10 4 ( ), 10 6 ( �9 .) or 10 8 (- ) N. caninum tachyzoites, or control inoculum ( " " ).

R e s u l t s

Experiment 1

Clinical observations. Sheep in group 1.1 (108 tachyzoites) showed a sharp rise in rectal temperature to reach a mean peak value of 41"2~ on day 4 (Fig. 1). Mean values declined to preinjection temperatures by day 9. Three of the six sheep showed a biphasic response with a decline (0"6~176 to temperatures of between 39-4~ and 39"8~ on day 6 or 7, followed the next day by a rise (0-6~176 In group 1.2 (106 tachyzoites), the mean temperature rose after day 6, reached a peak value of 40"8~ on day 8 and returned to a preinjection value by day 9. Three of the five animals showed a biphasic response with a decline (0-5~176 to individual temperatures of 39"3~ 39.5~ and 39"5~ on days 3, 6 and 9, respectively, followed by rises of 0"5~ 1-1~ and 0"7~ respectively. In group 1.3 (10 4 tachyzoites) there was evidence of a low mean peak of 40"3~ on day 9. Control animals in group 1.4 showed no rise in rectal temperature.

Serology. In group 1.1 the geometric mean of the IFAT IgG antibody titre to 2(. caninum had risen to 4096 by 4 weeks after injection of tachyzoites, but declined thereafter to 1024 by week 14 (Fig. 2). Mean titres in groups 1.2 and 1.3 rose more slowly and reached peaks just greater than and less than 2048, respectively, at week 5, before declining to mean values of 512 and 256, respectively, by week 22. After this, titres in all three groups rose steadily to geometric means of almost 8192 on week 34 (group 1.1) and 2048 between weeks 34 and 4,2 (groups 1,2 and 1.3) and remained hgh until the end of the study. Titres in group 1.4 were negative.

6 D . B u x t o n e t al.

Fig. 2.

8192 t

~o~o ~ ~_,~,.J, ~ ,,-"\'-,,T .-L--T--S"--~ 1024 l'Y

...., .T- , 0 10 20 30 40 50 60

Weeks after inoculation

Geometric mean titres and standard errors of IgG antibody to N. caninum of the four groups of sheep in experiment 1 given subcutaneous doses of 104 ( - - ) , 106 ( . . . . ) or 108 ( ) N. caninum tachyzoites, or control inoculum ( - - . ' ).

41 [ 4 o �84

~ i ~ ~ ?

Fig. 3.

38 ] ] ] I ] ] ] ] ] I ] ] I ] ] - 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Days af ter inoculation

Mean rectal temperatures and standard errors of pregnant sheep in experiment 2 after subcutaneous injection ofN. caninum tachyzoites ( - - - - - ) or control inoculum ( ).

Experiment 2

Clinical observations. After injection of tachyzoites, nine of the 12 ewes in group 2.1 showed a brief febrile response (>40"0~ between days 7 and 10 (Fig. 3), similar to that seen in group 1.2 in experiment 1. Three of the nine showed a biphasic response with a reduction in fever on day 6 or 7, followed the next day by a rise of 0-4~ "2~ Control ewes (group 2.2) did not show a febrile response.

Necropsy. At necropsy, the 12 infected ewes in group 2.1 were found to have 18 fetuses. All were alive when the dam was killed, except one, a mummified fetus, which was accompanied by a live sibling at 40 days post-inoculation

Infection of Sheep with Neospora caninum 7

512 -

25~

12s

64

3 2 -

1 6 -

8 -

+

I I I I

+ I I

§ 0

0 8 0 t l

CD II I I 1(33D

CD Q30 0 §

Fig. 4.

25 40 53

Days af ter inoculat ion

IgM antibody titres to N. caninum in the sera of fetuses in experiment 2 after inoculation of their dams with N. caninum tachyzoites (+) or control inoculum (O).

512 -

256

128

64

3 2 -

16

8

+

II III

III

I I

I l l

,~ ICD

(I?O

Fig. 5.

25 40 53

Days af ter inoculat ion

IgG antibody titrcs to N. caninum in the sera of fetuses in experiment 2 after inoculation of their dams with N. caninum tachyzoites (+) or control inocnlum (O).

(dpi). The eight control ewes in group 2.2 had a total of 15 live fetuses. There was no significant difference in weight of fetus, prescapular lymph node or spleen when infected and control animals were compared at any given sampling day.

Fetal serology. In fetuses from ewes in group 2.1, IgM and IgG antibodies to the parasite were detected by the IFAT at 25, 40 and 53 dpi (Figs 4 and 5). IgM values were highest at 25 dpi (256 and 512) and significantly higher (P<0"01) than control titres. At 40dpi, titres (16 to 128) were lower, but significantly higher than control titres (P<0"05); at 53 dpi they were lower still

8 D. B u x t o n e t a l .

+ a

1024

256

64

16

-1

/r /"

r

. . . . . . . .

I I I I I I 1 2 3 4 5 6 7 8

Weeks a l ~ r inoculation

Fig. 6. Geometric mean IgG antibody titres and standard errors to N. caninum in the ewes in experiment 2 after subcutaneous inoculation with N. can• tachyzoites ( . . . . . ) or control inoculum ( - - ) .

Table 1 The f requency of focal necrosis in placentomes from ewes experimental ly infected with N.

c a n i n u m at 90 days gestation (group 2.1)

Days after inoculation Mean number of Joci per Mean area of placentome " Foci of necrosis per 1000 mm ~ placentome • standard error (mm 2) -t- standard error

25 1"9• 590"6• 19"6 3"2 40 18"3_+4"9 519"1 +30"5 34"8* 53 2-9 • 0"4 511"3_+ 15"4 5"7

*Significantly more than either of the other two groups; P = <0"001 by analysis of variance. No lesions were found in any of the placental tissues from control animals.

(8 tO 128) and similar to control titres. IgG titres at 25 dpi (64 to 128) and 40dpi (128 to 512) were significantly higher than control titres (P<0"05 to <0.01). At 53 dpi, IgG titres (64 to 256) were significantly higher than control values (P<0"001). Sera from fetuses in group 2.2 consistently had IgM and IgG titres of <16 at 25 and 40dpi. At 53dpi, all control fetuses had IgM titres of _<32 except for one (64). Six had IgG titres of 8 or 16 and two had IgG titres of 64. No IgM or IgG antibody to N. caninum was detected in any of the fetal CSF samples.

Maternal serology. By the end of the second week after infection the geometric mean IgG titre was 128 (Fig. 6); it rose to a peak titre of almost 1024 at the fourth week before declining to a value of between 256 and 512 at 8 weeks dpi. No IgM or IgG antibody to N. caninum was detected in any of the maternal CSF samples.

Histopathological findings. Placenta. At 25 dpi, 0 10 small foci of necrosis were detected in each of the

three placentomes sampled from each fetus (Table 1). Each focus affected

Infection of Sheep with Neospora caninurrt 9

Fig. 7. A focus of necrosis in the placentome of a ewe (group 2.1) inoculated with tachyzoites of N. caninum at 90 days gestation and sampled at 25 dpi. The dark areas (arrows) contain cell debris in proteinaceous material, staining red, suggesting the presence of fibrin. V ~ fetal placental villous; S=ma te rna l caruncular septum. MSB. x 150.

fetal villous tissue and adjacent caruncular septum; between the two, aggregates of necrotic, strongly eosinophilic cells were seen, apparently composed, in part at least, of desquamated fetal trophoblast cells. At 40 dpi, foci of necrosis were more common (1-53 per placentome) (Table 1). The foci were also larger than at 25 dpi, and frequently the villous-septal space contained material which had the staining characteristics of serum and was usually shown by the MSB method to contain fibrin (Fig. 7). There was evidence of a mild fetal and maternal inflammatory response, with a local increase in the cellularity of fetal villous connective tissue and occasional lymphocytes associated with capillaries in affected areas of maternal caruncular septa.

At 53 dpi, the foci were similar in character but smaller than at 40 dpi and less numerous (0 to 10 per placentome; Table 1). Closely associated with the foci were signs of mild inflammation, with at tachment of mononuclear cells (chiefly lymphocytes) to the luminal surface of vascular endothelium in maternal caruncular septa and fetal villous capillaries; this was sometimes associated with mild hypercellularity in fetal villous connective tissue. All control placental tissues appeared normal.

Fetal brain and spinal cord. At 25 dpi there was a mild focal lymphoid meningitis, chiefly over the cerebrum. Sporadically scattered throughout the brain and spinal cord were blood vessels with pericyte proliferation and associated microgliosis (Table 2). Some other blood vessels were lightly cuffed by lymphoid cells. There were also partly organized foci in the neuropil, with palisading by microglia and lymphoid cells around a central necrotic core. Similar changes were present at 40dpi (Fig. 8), consistently in the cerebrum but

1 0 D . B u x t o n et al.

Table 2 D i s t r i b u t i o n Of h i s t o p a t h o l o g i c a l s i g n s o f i n f l a m m a t i o n and of n e o s p o r a an t i gen in the centra l

n e r v o u s s y s t e m (CNS) o f f e t u s e s f rom e w e s i n o c u l a t e d w i t h Neospora caninurn

Days after Fetus no. Histopathological lesions* in Neospora inoculation antigen in

cerebrum midbrain cerebellum medulla spinal cord CNS

25

40

53

1 + + _+ + + P A

2 + + + + + + T S

3 + + + + + + + + T S

4]" + - _+ - - P A

4]" + + + + + _+ T S

5 + -- -- + -- --

6]" + . . . . .

6]" . . . . . . . . . . . . . . . . . .

( m u m m y )

7 + + + + + + + + + + P A

8 + + + + + -- --

9]" _ + . . . .

9]" + -- -- + -- P A

lO]" + + - + - - lm + - - + - - l i t + . . . . . l l ]" + -I- - - - -

127 + - _+ - - - 12]. + -- _+ -- -- P A

* I n f l a m m a t i o n , i n c l u d i n g focal l y m p h o i d m e n i n g i t i s , focal m ic rog l i o s i s (with o r w i t h o u t c en t r a l necrosis) a n d v a s c u l a r cuffing.

~" D e n o t e s tw in fetuses.

+ + , M o d e r a t e ; + , mi ld ; • a t race; - , none ; PA, p a r t i c u l a t e a n t i g e n ; T S , t achyzo i t e - l ike s t ruc tures ;

. . . , n o t d o n e because fe tus m u m m i f i e d .

N o i n f l a m m a t i o n was f o u n d in the b r a i n s o f c on t ro l fetuses.

inconsistently elsewhere. At 53 dpi, lesions were less common and largely confined to the cerebrum, with focal microgliosis, vascular cuffing (Fig. 9) and partly organized inflammatory foci with central necrosis (Table 2). No abnormalities were found in any of the fetuses from control ewes.

Other fetal tissues. At 25 dpi, in four of five fetal prescapular lymph nodes from group 2.1, the cortex was readily discernible and contained primary nodules and a few secondary follicles. By comparison, in similar nodes from group 2.2 cortical development was much less advanced, although there was no statistically significant difference in the number of follicles between the two groups (Table 3). At 40 dpi, in three of four fetal prescapular nodes from group 2.1, follicular development was more advanced than in group 2.2; and at 53 dpi, seven of eight nodes in group 2.1 showed reactive changes, with medullary cords and sinuses both well formed. At 40 dpi there were significantly more primary and secondary follicles in fetuses from group 2.1 than in those from group 2.2, while at 53 dpi there were significantly more primary follicles in group 2.2 and significantly more secondary follicles in group 2.1 when compared with each other (Table 3). In all fetuses in group 2.1 sampled at 25 dpi there was mild to moderate diffuse myocarditis and associated focal endocarditis and pericarditis; at 40 and 53 dpi, mild lesions were seen in the

Infection of Sheep with Neospora caninum 1 1

Fig. 8. A characteristic focus of necrosis surrounded by inflammatory cells in the cerebrum of a fetus taken from a ewe (group 2.1) inoculated with tachyzoites ofN. caninum at 90 days gestation and sampled at 25 dpi. HE. x 150.

Fig. 9. Focal microgliosis and perivascular cuffing by mononuclear cells in the cerebrum of a fetus taken from a ewe (group 2.1) inoculated with tachyzoites ofN. caninum at 90 days gestation and sampled a t53dp i . HE. x150.

hearts of three of four and two of seven fetuses examined, respectively. In the lung, very mild focal inflammation, consisting of scattered small aggregates of mononuclear cells, was present in three of five, one of four and five of

12 D. B u x t o n e t al.

Table 3 A comparison of the number o f primary and secondary fol l ic les in the prescapular lymph

nodes o f fetuses from infected (group 2.1) and control (group 2.2) ewes in exper iment 2

Group Days after inoculation Number of primaF Number of secondary follicles (+_ se) follicles (+_ se)

2.1 25 3-60 (0-93) 1-60 (0-81) 2.2 25 3.75 (1.65) 0.25 (0.25) 2.1 40 10.0 (1.87)* 7.50 (0.65)~ 2.2 40 3-60 (1-20)* 2-66 (1-76)]- 2.1 53 4.25 (1.22) + 7.25 (2.02)w 2.2 53 12.0 (1.07)~ 0.63 (0-37)w

*P= <0"05; ~'P= 0"05; ~.P= <0"001 ; w 0"01.

Fig. 10. The discrete black foci (small arrows) represent the immunohistochemical demonstration of N. caninum tachyzoites and the larger dark focus (large arrow) represents antigenic material in the placentome of a ewe (group 2.1) inoculated with tachyzoites ofN. caninum at 90 days gestation and sampled at 40 dpi. V = fetal placental villus; S = maternal caruncular septum. Immunoperoxidase. x 375.

eight fetuses examined at 25, 40 and 53 dpi, respectively. In the liver, light accumulations of periportal lymphoid cells were present in all fetuses from group 2.1, with the exception of one at 40dpi and one at 53dpi. Focal hepatocyte necrosis was present in two at 25 dpi and one at 53 dpi. The spleen, thymus, digestive tract and kidney appeared normal in all fetuses. No significant lesions were detected in any of the fetuses from group 2.2.

Maternal tissues. HE-stained sections of all maternal tissues sampled appeared normal and no evidence of N. caninum was found.

Immunohistochemistry. Neospora antigen, associated with placental focal necrosis, was present in all animals inoculated with N. caninum but was not abundant; it took the form of tachyzoite-like structures as well as cellular debris (Fig.

Infect ion o f S h e e p w i t h Neospora caninum 13

Fig. 11. A small, mild focus of microgliosis in the brain of a fetus taken from a ewe inoculated with tachyzoites of N. caninum at 90 days gestation and sampled at 25 dpi. Note the small discrete black focus (arrow), which represents a closely packed aggregation of N. caninum tachyzoites. Immunoperoxidase. x 375.

10). Neospora antigen was observed in the brains of all fetuses from infected ewes sampled at 25 dpi. Antigen was detected in the brains of only two other fetuses, both of which were sampled at 53 dpi. Positive immunohistochemical labelling was always associated with focal inflammation and was frequently located within the central necrotic tissue. In all cases granular material was positive and in three fetuses at 25 dpi positive tachyzoite-like structures were also present (Fig. 11). The latter were also detected in the heart of one fetus from an infected ewe at 25 dpi, and antigen-containing particles were present in a focus of hepatic necrosis in a fetus from another ewe sampled at the same time.

D i s c u s s i o n

Experiment 1 showed that sheep could be infected readily with N. caninum tachyzoites injected sc and that a graded response was obtained according to the infective dose; a readily measurable fever followed the injection of 108 tachyzoites, but doses of 106 and 10 a produced reduced responses. The febrile reaction in some sheep was biphasic, but an explanation of this phenomenon will require further research, as will the observation that after the injection of N. caninum tachyzoites antibody titres peaked and then declined at about week 5 or 8, only to rise again at week 22. All sheep were seropositive after 1 year, showing that, as in infections with T. gondii, serum antibody can persist for a long period. Buxton et al. (1993) noted, however, that an exception was provided by sheep infected with "incomplete" $48 T. gondii tachyzoites, in

14 D. Buxton e t al.

which, in the absence of tissue cyst development, antibody fell to low titres. This observation gives rise to speculation that N. caninum tissue cysts may have persisted in the tissues of the sheep in this experiment, albeit that none was detected.

The intermediate dose of 10 6 tachyzoites was selected for experiment 2. The febrile and serological responses observed were similar to those in experiment 1, indicating that the viability and virulence of the inocula remained constant and that the animals underwent a primary response to the parasite. From this it can be concluded that an IgG IFAT titre of _<64 in an adult sheep is an indication that it has not been exposed to N. caninum.

One fetus found dead and mummified at necropsy was considered to have died as a result of the neospora infection. The other 17 fetuses from infected ewes were alive just before necropsy, but all showed lesions, most commonly of the placentas and CNS. McAllister et al. (1996) found neospora tissue cysts only in the CNS ~tnd observed tachyzoites in HE-stained sections of placental tissues. In the current study, with the exception that neospora antigen was demonstrated in one section of liver and one of heart at 25 dpi, the results were broadly similar, the parasite and its antigens being demonstrable in CNS and placenta. In the latter it was associated with foci of necrosis, which were larger and more numerous at 40 dpi than at 25 or 53 dpi. The presence of serum leakage suggested vascular damage, but whether this was caused directly or indirectly by N. caninum is not known. Within the limits of this experiment the parasite-induced placental damage appeared to be progressive up to c. 40 dpi, after which the lesions started to regress.

By comparison with placental lesions, CNS lesions were more widespread at 25 dpi, when neospora antigen was most readily demonstrated; they had become less widespread at 40 dpi and were largely confined to the cerebrum at 53 dpi. Thus, while placental lesions due to N. caninum had been initiated by 25 dpi and were becoming progressively severe up to 40 dpi, the parasite had already invaded the developing fetus at 25 dpi (115 days' gestation) and caused widespread lesions, which were resolving by 40 dpi and 53 dpi (130 and 143 days' gestation, respectively). This sequence of events was presumably influenced by a number of factors, including dose and virulence of the parasite and fetal immune response, which itself was dependent on fetal age.

Detection of antibody in fetal serum provided evidence of a specific fetal immune response. Both IgM and IgG antibodies to the parasite were detected at 25 dpi, with lower IgM titres but higher IgG titres at 40 and 53 dpi. Of the control fetuses sampled at 53 dpi, one had an IgM titre of 64 and two had IgG titres of 64. However, from the data shown in Figs 4 and 5 it is concluded that a titre of _> 64 would be expected to indicate fetal exposure to N. caninum. The humoral response in infected fetuses was associated with increased follicular development in the lymph nodes examined and it is probable that cell-mediated immune mechanisms are also important in the ruminant response to N. caninum (Innes et al., 1995). The presence of lymphoid inflammatory lesions in infected fetuses supports this view.

In the CNS, a cellular immune response with both lymphoid and microglial

Infection of Sheep with Neospora caninum 15

inflammation was detected. The neuropathological changes, which char- acteristically included foci of necrosis and surrounding inflammatory response, closely resembled those seen by others in cases of experimental ovine neo- sporosis (Dubey and Lindsay, 1990; McAllister et al., 1996) and in experimental and natural ovine toxoplasmosis (Hartley and Kater, 1963; Buxton and Finlayson, 1986; Buxton, 1989). In addition, they are similar to descriptions of lesions seen in natural (Thi|sted and Dubey, 1989; Barr et al., 1990; Thornton et al., 1991; Wouda et al., 1992) and experimental (Barr et al., 1994) bovine neosporosis.

This study demonstrated that N. caninum can infect sheep readily and be pathogenic for the ovine fetus, and that the resulting lesions may resemble those induced by T. gondii in pregnant ewes. It is important, therefore, that diagnostic laborat~0ries should attempt to distinguish these two infections. In addition, this Study demonstrated that experimental neospora infection of pregnant and non-pregnant sheep provides a good model of bovine neosporosis.

Acknowledgments

We thank Mr John Spence for his assistance in measuring the areas of the placentomes and Dr Hugh Reid for helpful discussions about the manuscript. This study was funded by the Scottish Office Agriculture, Environment and Fisheries Department.

References

Barber, J. S., Holmdahl, O . J .M. , Owen, M. R., Guy, F., Uggla, A. and Trees, A. J. (1995). Characterization of the first European isolate ofNeospora caninum (Dubey, Carpenter, Speer, Topper and Uggla). Parasitology, 111, 563-568.

Barlow, R. M. (1972). Experiments in Border disease IV. Pathological changes in ewes. Journal of Comparative Pathology, 83, 151-157.

Barr, B. C., Anderson, M. L., Blanchard, P. C., Daft, B. M., Kinde, H. and Conrad, P. A. (1990). Bovine fetal encephalitis and myocarditis associated with protozoal infections. Veterinary Pathology, 27, 354-361.

Barr, B. C., Anderson, M. L., Woods, L. W., Dubey, J. P. and Conrad, P. A. (1992). Neospora-like protozoal infections associated with abortion in goats. Journal of Veterinary Diagnosis and Investigation, 4, 365 367.

Barr, B. C., Rowe, J. D., Sverlow, K. W., BonDurant, R. H., Ardans, A. A., Oliver, M. N. and Conrad, P. A. (1994). Experimental reproduction of bovine fetal Neospora infection and death with a bovine Neospora isolate. Journal of Veterinary Diagnosis and Investigation, 6, 207-215.

Bjerkas, I., Mohn, S. F. and Presthus, J. (1984). Unidentified cyst-forming sporozoan causing encephalomyelitis and myositis in dogs. Zeitschriftfiir Paras#enkunde, 70, 271-274.

Buxton, D. (1989). Toxoplasmosis in sheep and other farm animals. In Practice, 11, 9-12.

Buxton, D., Blewett, D. A., Trees, A. J., McColgan, C. and Finlayson, F. (1988). Further studies in the use of monensin in the control of experimental ovine toxoplasmosis. Journal of Comparative Pathology, 98, 225-236.

Buxton, D. and Finlayson, J. (1986). Experimental infection of pregnant sheep with Toxoplasma gondii: pathological and immunological observations on the placenta and foetus. Journal of Comparative Pathology, 96, 319-333.

Buxton, D., Maley, S. W., Thomson, K. M., Innes, E. A. and Trees, A. J. (1995). Experimental infection of pregnant sheep with Neospora caninum. Proceedings of the 13th Congress on Veterinary Pathology, Edinburgh, 27 30 September, 2A-6.

16 D. Buxton e t al.

Buxton, D., Thomson, K. M., Maley, S., Wright, S. and Bos, H.J. (1993). Experimental challenge of sheep 18 months after vaccination with a live ($48) Toxoplasma gondii vaccine. Veterinary Record, 133, 310-312.

Dubey, J. P., Acland, H. M. and Hamir, A. N. (1992). Neospora caninurn (Apicomplexa) in a stillborn goat. Journal of Parasitology, 78, 532-534.

Dubey, J. P., Carpenter, J. L., Speer, C. A., Topper, M.J. and Uggla, A. (1988). Newly recognised fatal protozoan disease of dogs. Journal of the American Veterinary Medical Association, 192, 1269 1285.

Dubey, J. P., Hartley, W.J., Lindsay, D. S. and Topper, M.J. (1990). Fatal congenital Neospora caninum infection in a lamb. Journal of Parasitology, 76, 127 t30.

Dubey, J. P. and Lindsay, D. S. (1990). Neospora caninum-induced abortion in sheep. Journal of Veterinary Diagnosis and Investigation, 2, 230 233.

Dubey, J. P., Morales, J. A., Villalobos, P., Lindsay, D. S., Blagburn, B. L. and Topper, M.J. (1996). Neosporosis-associated abortion in a dairy goat. Journal of the American Veterinary Medical Association, 208, 263 265.

Dubey, J. P. and Porterfield, M. L. (1990). Neospora caninum (Apicomplexa) in an aborted equine fetus. Journal of Parasitology, 76, 732 734.

Ellis, J., Luton, K., Baverstock, P. R., Brindley, P.J., Nimmo, K. A. and Johnson, A. M. (1994). The phylogeny of Neospora caninum. Molecular and Biochemical Para- sitology, 64, 303-311.

Hartley, W.J. and Kater, J. C. (1963). The pathology of Toxoplasma infection in the pregnant ewe. Research in Veterinary Science, 4, 326 332.

Holmdahl, O . J . M . , Mattsson, J. G., Uggla, A. and Johansson, K.-E. (1994). The phylogeny ofNeospora caninum and Toxoplasma gondii based on ribosomal sequences. FEMS Microbiology Letters, 119, 187-192.

Innes, E. A., Panton, W. R. M., Marks, J., Trees, A. J., Holmdahl, J. and Buxton, D. (1995). Interferon gamma inhibits the intraeellular multiplication of Neospora caninum, as shown by incorporation of 3H uracil. Journal of Comparative Pathology, 113, 95-100.

Lindsay, D. S. and Dubey, J. P. (1989). In vitro development of Neospora caninum (Protozoa: Apicomplexa) from dogs. Journal of Parasitology, 75, 163-165.

McAllister, M. M., McGuire, A. M., Jolley, W. R., Lindsay, D. S., Trees, A.J. and Stobart, R. H. (1996). Experimentally induced neosporosis in pregnant ewes and their offspring. Veterinary Pathology, 33, 647 655.

Thilsted, J. P. and Dubey, J. P. (1989). Neosporosis-like abortions in a herd of dairy cattle. Journal of Veterinary Diagnosis and Investigation, 1,205-209.

Thornton, R. N., Thompson, E.J. and Dubey, J. P. (1991). Neospora abortion in New Zealand cattle. New Zealand Veterinary Journal, 39, 129-133.

Woods, L. W., Anderson, M. L., Swift, P. K. and Sverlow, K. W. (1994). Systemic neosporosis in a California black-tailed deer (Odocoileus hemionus columbianus). Journal of Veterinary Diagnosis and Investigation, 6, 508-510.

Wouda, W., van den Ingh, T. S. G. A. M., van Knapen, F., Sluyter, F.J.H., Koeman, J. P. and Dubey, J. P. (1992). Neospora abortus bij bet rund in Nederland. Tijdschrift voor Diergeneeskunde, 117, 599 602.

I Received, November 15th, 1996] Accepted, March 6th, 1997 J